lemma
proved
phi_differentiable
show as:
view math explainer →
open explainer
Read the cached plain-language explainer.
open lean source
IndisputableMonolith.Cost.AczelProof on GitHub at line 41.
browse module
All declarations in this module, on Recognition.
explainer page
depends on
used by
formal source
38 intervalIntegral.integral_hasDerivAt_right (h_cont.intervalIntegrable 0 t)
39 h_cont.aestronglyMeasurable.stronglyMeasurableAtFilter h_cont.continuousAt
40
41private lemma phi_differentiable (H : ℝ → ℝ) (h_cont : Continuous H) :
42 Differentiable ℝ (Phi H) :=
43 fun t => (phi_hasDerivAt H h_cont t).differentiableAt
44
45private lemma deriv_phi_eq (H : ℝ → ℝ) (h_cont : Continuous H) : deriv (Phi H) = H :=
46 funext fun t => (phi_hasDerivAt H h_cont t).deriv
47
48private lemma exists_integral_ne_zero (H : ℝ → ℝ) (h_one : H 0 = 1) (h_cont : Continuous H) :
49 ∃ δ : ℝ, 0 < δ ∧ Phi H δ ≠ 0 := by
50 have h_pos : (0 : ℝ) < H 0 := by rw [h_one]; exact one_pos
51 have h_ev : ∀ᶠ x in nhds (0 : ℝ), (0 : ℝ) < H x :=
52 h_cont.continuousAt.eventually (Ioi_mem_nhds h_pos)
53 obtain ⟨ε, hε_pos, hε⟩ := Metric.eventually_nhds_iff.mp h_ev
54 refine ⟨ε / 2, by positivity, ?_⟩
55 intro h_eq
56 have hδ_pos : (0 : ℝ) < ε / 2 := by positivity
57 obtain ⟨c, hc_mem, hc_eq⟩ := exists_hasDerivAt_eq_slope (Phi H) H hδ_pos
58 ((phi_differentiable H h_cont).continuous.continuousOn)
59 (fun x _ => phi_hasDerivAt H h_cont x)
60 rw [phi_zero, h_eq, sub_zero, zero_div] at hc_eq
61 linarith [hε (show dist c 0 < ε by
62 simp only [Real.dist_eq, sub_zero, abs_of_pos hc_mem.1]; linarith [hc_mem.2])]
63
64private lemma representation_formula (H : ℝ → ℝ) (h_cont : Continuous H)
65 (h_dAl : ∀ t u, H (t + u) + H (t - u) = 2 * H t * H u)
66 {δ : ℝ} (hδ_ne : Phi H δ ≠ 0) (t : ℝ) :
67 H t = (Phi H (t + δ) - Phi H (t - δ)) / (2 * Phi H δ) := by
68 -- Strategy: avoid integral substitution by using FTC + is_const_of_deriv_eq_zero.
69 -- We prove ∫₀ᵈ H(t+u)du = Phi(t+d)-Phi(t) and ∫₀ᵈ H(t-u)du = Phi(t)-Phi(t-d)
70 -- by showing both sides have the same derivative (w.r.t. d) and agree at d=0.
71 -- Then the d'Alembert equation integrated gives the representation.